Application of Variational AutoEncoder (VAE) Model and Image Processing Approaches in Game Design.

In recent decades, the Variational AutoEncoder (VAE) model has shown good potential and capability in image generation and dimensionality reduction. The combination of VAE and various machine learning frameworks has also worked effectively in different daily life applications, however its possible use and effectiveness in modern game design has seldom been explored nor assessed. The use of its feature extractor for data clustering has also been minimally discussed in the literature neither. This study first attempts to explore different mathematical properties of the VAE model, in particular, the theoretical framework of the encoding and decoding processes, the possible achievable lower bound and loss functions of different applications; then applies the established VAE model to generate new game levels based on two well-known game settings; and to validate the effectiveness of its data clustering mechanism with the aid of the Modified National Institute of Standards and Technology (MNIST) database. Respective statistical metrics and assessments are also utilized to evaluate the performance of the proposed VAE model in aforementioned case studies. Based on the statistical and graphical results, several potential deficiencies, for example, difficulties in handling high-dimensional and vast datasets, as well as insufficient clarity of outputs are discussed; then measures of future enhancement, such as tokenization and the combination of VAE and GAN models, are also outlined. Hopefully, this can ultimately maximize the strengths and advantages of VAE for future game design tasks and relevant industrial missions.

Spatial and Socio-Classification of Traffic Pollutant Emissions and Associated Mortality Rates in High-Density Hong Kong via Improved Data Analytic Approaches.

Excessive traffic pollutant emissions in high-density cities result in thermal discomfort and are associated with devastating health impacts. In this study, an improved data analytic framework that combines geo-processing techniques, social habits of local citizens like traffic patterns and working schedule and district-wise building morphologies was established to retrieve street-level traffic NOx and PM2.5 emissions in all 18 districts of Hong Kong. The identification of possible human activity regions further visualizes the intersection between emission sources and human mobility. The updated spatial distribution of traffic emission could serve as good indicators for better air quality management, as well as the planning of social infrastructures in the neighborhood environment. Further, geo-processed traffic emission figures can systematically be distributed to respective districts via mathematical means, while the correlations of NOx and mortality within different case studies range from 0.371 to 0.783, while varying from 0.509 to 0.754 for PM2.5, with some assumptions imposed in our study. Outlying districts and good practices of maintaining an environmentally friendly transportation network were also identified and analyzed via statistical means. This newly developed data-driven framework of allocating and quantifying traffic emission could possibly be extended to other dense and heavily polluted cities, with the aim of enhancing health monitoring campaigns and relevant policy implementations.